Export a 24-bit WAV at your session's sample rate, mastered to roughly -14 LUFS integrated with a true peak ceiling of -1 dBTP, and fill in the metadata completely. Every streaming service transcodes what you send into its own formats, so your job is to hand over the cleanest possible source and leave enough headroom that the transcode does not clip.
- A finished mix with headroom on the master bus
- A loudness meter that reads integrated LUFS and true peak
- Your final metadata: track title, artist, ISRC if you have one, contributor credits
- Your distributor's specific requirements, which override anything general
- Bit depth: 24-bit. There is no reason to export 16-bit. If a distributor requires it, dither on the way down.
- Sample rate: whatever you recorded at. 44.1kHz and 48kHz are both fine. Do not upsample — it adds nothing. Do not downsample unless you have to, and if you do, use a good sample rate converter.
- Format: WAV or FLAC. Both are lossless. AIFF is fine too.
- Never MP3. Uploading a lossy file means the service transcodes a transcode, and the artifacts compound.
Streaming services measure the loudness of your track and adjust playback so it sits at a consistent level alongside everything else. If your master is louder than their target, it gets turned down. It does not get to be louder than the next track.
The practical consequence: squashing a master to -6 LUFS gains you nothing on streaming and costs you dynamics. It comes out at the same perceived level as a -14 LUFS master, but with less punch and less depth.
Aim for around -14 LUFS integrated as a default reference. Then apply judgment:
- Quiet, dynamic material — folk, jazz, classical — often sits happily at -16 to -18 and should not be pushed
- Dense electronic and hip hop frequently sit at -9 to -11 and still sound correct, because the material has less dynamic range to lose
- The right answer is the one where the master sounds good, not the one where the number is exact
Measure integrated loudness over the whole track, not short-term.
Set your limiter's ceiling to -1 dBTP and make sure the meter is reading true peak, not sample peak.
The reason is inter-sample peaks. A waveform that never exceeds 0 dBFS at the sample points can still exceed it between them, and lossy encoders reconstruct those points. A master that reads -0.1 dBFS sample peak can clip audibly after transcoding. One decibel of headroom costs nothing.
- Trim silence at the start to a few milliseconds, not several seconds
- Apply a short fade at the very start and end to avoid clicks, even on tracks that end cold
- Leave any intentional reverb tail intact — do not cut it short to save file size
- Check the whole file end to end at least once before uploading
| Field | Notes |
|---|
| Track title | Exactly as it should display, including features |
| Artist name | Consistent across every release — this drives your artist page |
| Album or single title | Match the release |
| ISRC | One unique code per recording; your distributor can issue them |
| Songwriter credits | Required for royalties to route correctly |
| Producer and engineer credits | Increasingly displayed and worth filling in |
| Explicit flag | Set it accurately |
Metadata errors are the single most common reason a release gets delayed or a payment goes missing.
Play the exported file itself — not the session — start to finish, on at least two systems. Check for clipping, a mislabelled version, a plugin that was bypassed during bounce, and count-ins that made it into the file.
| Setting | Value |
|---|
| Format | WAV or FLAC |
| Bit depth | 24-bit |
| Sample rate | Session rate, 44.1 or 48kHz |
| Integrated loudness | About -14 LUFS, genre-adjusted |
| True peak ceiling | -1 dBTP |
| Dither | Only if going to 16-bit |
Related reading: LUFS and loudness explained, what mastering is for beginners, and how to distribute music.
Start making music in Veena
Free, browser-based, no downloads required.
Try Veena Free